The Effect of Oxygen on the Fluorescence of Certain Hydrocarbons'

J. A. MILLER AND C. A. BAUMA". VOI. 65. (d) Isomerization by Melting.-A sealed tube contain- ing 20 mg. of the carotenoid was maintained at 140-145' f...
0 downloads 0 Views 1MB Size
J. A. MILLERAND C. A. BAUMA"

1540

(d) Isomerization by Melting.-A sealed tube containing 20 mg. of the carotenoid was maintained at 140-145' for five minutes. After plunging into ice water, the solidified mass was taken up in cold benzene and chromatographed 1'8 X 7 cm.), following dilution with petroleum ether. 100 colorless t o D section

66

pink: gazakaxanthin: 494.5, 462.5, 434.5 ing (with iodine, 490.5, 458.5 rnp) ,?I brownish pink: 489, 457 m p (490.5, 459 mp) 3 orange brown: 488, 456.5 m p (491, 458.5 mp) 15 orange brown: 485. 451.5 mu 1490.5, 458.5 mu) 12 yelloyv: 484,453 6 rnp (491 5, 459 mpii 15 yellow: 486, 454 ;inip (491, 458.5 mfi) 12 pink: 489.3, 457 5 rnp (492, 459 tnp) 1 traces of color 191,4RU tinip (491, 459 mp)

Summary 1. Gazaniaxanthir~,~ the main pigment of Gazunia rigens flowers, has been studied by means

iCONTRIBUTIOPi FROM THE

VOI. 65

of methods causing trans-cis changes in solution and in melt. On the Tswett column two groups of neo-compounds appeared, the further differentiation of which was difficult. 2. The molecular extinction curve of all-trunsgazaniaxanthin shows a very flat maximum around 350 mp. Upon heating or iodine catalysis, however, a marked maximum develops in this region ("cis-peak") .fi 3 . Catalytic hydrogenation indicates that gazaniaxanthin may be dihydrorubixanthin. 4. The carotenoid mixture in Gazania flowers grown in Portugal3 differs from that occurring in flowers grown in Southern California. PASADENA, CALIFORNIA

RECEIVED MARCH4,1943

DEPARTMENT O F BIOCHEMISTRY, COLLEGE O F AGRICULTURE, UNIV&RSITY O F WISCONWN]

The Effect of Oxygen on the Fluorescence of Certain Hydrocarbons' BY J. A. MILLERAND C,. A. RAUMANN Oxygen is known to diminish the fluorescence of many corn pound^^,^^*^^ including certain poly-

cyclic h y d r o c a r b ~ n s . ~Weil-Malherbe ~~~~ and 'IVeiss2 observed that the fluorescence of benzpyrene was nearly the same in several solvents from which oxygen had been removed by suction, but that on oxygenation varying degrees of quenching resulted. The quenching effect was greater in oxygen than in air, and was completely arid rapidly reversible. Recently we have observed that the fluorescence of several carcinogenic hydrocarbons varies markedly with the solvent employed.6 The results did not, however, indicate to what extent the variations were due t.o dissolved oxygen or directly to the solvent itseli. Accordingly the effect of oxygen has been measured on the fluorescence of 3,4-benzpyrene (BPI, 20-methylcholanthrene (MC), 9,lO-dimethyl-1,2-benzanthracene (DMBA), 1,2,5,6-dibenzanthracene (DBA), 1,2-benzanthracene (BA), and anthracene (A), each in several representative solvents. Low concentrations of hydro(1) Published with t h e approval of t h e Director of t h e Wisconsin Agricultural Experiment Station. We are indebted to the Wisconsin Research Foundatiou and t o t h e Jonathan Bowman Cancer Research Fund for financial support. ( 2 ; \Veil-hfalherbe and Weiss. A'alura, 149, 471 (1942). ( 3 1 Rowen and Nortoti, 7'vans. Favuday Sac.. 3E, 44 (1939). Kdutsky. i?e7., :iiilcr an(! D'LU

carbon were employed. The solutions were placed in long all-glass fluorometer tubes as previously described,' and the dissolved air removed a t the pump. The intensity of fluorescence was measured in a Coleman photofl~orometer~ in vacuo* and again after shaking the solution with known mixtures of oxygen and nitrogen. Deductions were then made regarding the composition of a hydrocarbon-inhibitor complex assumed to be present in the quenched solutions. Other gases and other inhibitors of fluorescence were studied in a similar manner, and in certain cases the effect of the inhibitor on the hydrocarbon compared with its effect on the unsaponifiable matter of mouse tissues. General Results I n vacuo, the fluorescence of dilute solutions of benzpyrene was nearly the same in the six solvents employed (Table i). In air, however, marked variations with solvent were observed, and usually the readings were less than half of those observed in vacuo. This parallels the results of Weil-Malherbe and Weiss.2 However, in tetrahydrofurfuryl alcohol, the fluorescence observed in air was over SSY0 of that observed in vacuo, while in aniline or in carbon tetrachloride (7) hfiller a n d Baumann, ibid., 3,217 (1943). (8) I n this paper the phrase "rn Iacim' refers to t h e absence of all gas b u t solvent vapor

Aug., 1943

THE

EFFECTOF OXYGEN

ON THE

.

FLUORESCENCE OF CERTAIN HYDROCARBONS 1541

TABLE I THEFLUORESCENCE OF HYDROCARBONS IN VARIOUS SOLVENTS i n Vacuo AND Hydrocarbon, , PdCC.

BP . . . . . , , , , . . . . .0.02 vac.

IN

AIR

Fluorescence, Ammeter Units" DMBA

MC vac.

air

vac.

BA

DBA

2.00

0.10

0.10 air

air

vac.

.50

air

vac.

air

A .50 vac. air

56.7 11.5 47.9 12.0 30.0 8 . 5 51.0 33.0 60.6 9 . 3 33.0 6 . 3 Petroleum ether 88.4 27.5 120.6 1 9 . 5 45.7 11.5 57.2 41.0 63.0 17.2 58.9 15.8 Acetone 67.6 27.5 103.2 21.5 40.1 1 4 . 5 46.7 16.5 Propionic acid 61.5 20.0 50.8 43.0 75.0 36.0 43.6 17.5 59.6 2 7 . 0 113.6 29.0 50.2 23.5 52.8 41.0 n-Butyl alcohol 90.0 53.0 70.6 34.8 126.3 36.5 54.3 24.0 74.7 40.0 56.6 51.8 Dioxane 105.2 62.5 130.5 41.5 59.7 39.5 Pyridine 60.2 34.8 117.2 63.0 79.0 76.0 The intensity of the fluorescence was expressed in terms of the arbitrary 100 divisions of the ammeter scale, with the instrument set so that 0.0003% quinine sulfate gave a deflection of 50 units.'

fluorescence was negligible under both conditions. The fluorescence intensities of the other hydrocarbons were similarly depressed by oxygen, but the intensities in vacuo varied widely from one another for any given hydrocarbon (Table I). In every case the quenching was completely and rapidly reversible. The fluorescence of anthracene, the only compound studied lacking an angular benzene ring, was not so greatly affected by oxygen as the other hydrocarbons studied. A4ta relatively low partial pressure of oxygen as in air, the fluorescence of benzpyrene in petroleum ether was 15.5% of that in VUCUO, but for a further diminution much larger amounts of oxygen were required (Fig. 1). Thus fluorescence varied hyperbolically with oxygen tension. However, a change in air pressure of 10 mm. altered fluorescence by only 0.5 ammeter unit7 or 1% in a reading of 50, the limit of accuracy of our instrument. Changes in temperature affected the fluorescence of the hydrocarbons more markedly. The solubility of oxygen in liquids such as petroleum ether decreased with increase in temperature, and the fluorescence of benzpyrene increased correspondingly. At ordinary room temperatures an increase of 10' increased fluorescence by approximately 6%. However, the combined changes in temperature and pressure within a room appeared to be of only minor consequence in the fluorometric determination of any of the hydrocarbons, since very reproducible readings have been obtained during the past 18 months, even with solvents such as petroleum ether in which the solubility of oxygen is high. In vucuo, temperature extremes from 0 to 50' affected the fluorescence of benzpyrene in petroleum ether by less than * 1%. I n Table 11 representative solvents are listed according to their abilities to dissolve oxygen. Liquids such as petroleum ether, naphtha, ether

and carbon tetrachloride dissolved large amounts of oxygen, and in these solvents the fluorescence of benzpyrene was low in the presence of air.

300

100

300 500 700 Po**mm. Fig. 1.-The fluorescence of benzpyrene in petroleum ether (lgg./cc.) a t various partial pressures of oxygen. The circles indicate observed intensities of fluorescence; the curves are calculated curves for the possible combinations 2BP.02, BP.02, BP.202. The derived equations are for 2BP.Oz, F =

-nkl(l

*

+

dl 8KkPo, 4kKP0,

Each calculated curve passes through the point (F = nkl, Pol = 0), the observed fluorescence i n mcuu, and through the observed point chosen for the calculation of the KkY values, viz., the fluorescence in air. The calculated curve that agrees with the observed data is thus common t o all reliable observed points and is evidence for the presence of the assumed complex. However, a new family of curves for the other possible complexes is obtained for every point chosen for the evaluation of Kku.

J. A. MILLERAND C. A. BAUMANN

1542 TABLE I1

FLUORESCENCE OF BENZPYRENEAND SOLUBILITY OR OXKORN I N VARIOUS SOLVENTS Fluoreacenee. ammeter voitr.* 0.1 pg. BP/cr. at 25-C.

Solvent

Solubility of 01;

cc.

01 at 9. T. P., c c solvent a t I , 'C.b

Ethyl ether 95.0 0.415 at 20.3' Petroleum ether (UD to 46.5' .409 18.5 65°C..d. 0.608)' Petroleum ether 1 6 s 57.0d .292 18 100°C.. d. 0.7W) Carbon tetrachloride 8.0 .225 25 Naphtha (b. p. 57.0 .21+ 28 106-117') Acetone 86.0 ,207 19.3 Chloroform 115.0 ,205 16.3 Methanol 79.0 .175 18.8 Xylene 104.0 ,169 16 Toluene 105.0 ,168 18 Benzene 120.0 .I63 19 Ethyl acetate 77.0 ,163 20 Isoamyl alcohol 145.0 ,163 17.3 Ethanol, abs. 94.0 ,143 19.8 Pyridine 175.0 .12t 28 Pyridine 175.0 ,099 18.5 Ethanol, 95Y0 120.0 . I Z t 28 Dioxane 174.0 .12t 28 Methyl cellosolve 235.0 12t 28 Tetrahydrofurfury1 alcohol 286.0 .lot 28 Nitrobenzene 0.0 ,070 18.5 a Defined in Table I. The solubilities of oxygen are those recorded in the "International Critical Tables." 1st ed. 1928,Vol. 111, pp. 261-283,except the values markedt, which are approximatiom made in this Laboratory in B Van Slyke apparatus. Fluorescence reading in Skelly Fluorescence reading in naphtha. Solve B,b. p. 66-67". b. p. 106-117". ~~

~

'

Conversely, tetrahydrofurfuryl alcohol, pyridine, and dioxane dissolved lesser amounts of oxygen, and they permitted much greater intensities of fluorescence in the presence of air. Furthermore, within the series-benzene, toluene, xylenefluorescence varied inversely as the solubility of oxygen. This relationship, however, was not

K

i

A

Fig. Z.-Fluorcicenee

.-,&til

A.

spectra of benzpyrene in vacuum and in air.

VOl. 6.5

an exact one. Ethyl ether and petroleum ether absorbed nearly the same amounts of oxygen, yet the fluorescence of benzpyrene in the former was double that in the latter. Isoamyl alcohol is reported to dissolve as much oxygen as ethyl acetate, yet fluorescence in the two solvents was markedly different (Table 11). Nitrobenzene dissolved little oxygen, but it completely prevented the fluorescence of the hydrocarbon. Evidently solubility of oxygen was not the only factor through which solvents influenced the intensity of the fluorescence of dissolved hydrocarbons. The ether linkage appeared t o be one of these other factors: fluorescence was higher in tetrahydrofurfuryl alcohol, methyl cellosolve, dioxane, pyridine (N-analog of an ether), and diethyl ether than the small solubilities of oxygen mieht lead one to expect. Benzpyrene is probably less subject t o such solvent effect per se than the other hydrocarbons, for th? intensity of its fluoresccnce in vacuo was approximatelv the same for several solvents. Methylcholanthrene and the other carcinogens, on the contrary, yielded fluorescence intensities in vacuo which varied widely from solvent to solvent. These differences must have been due to the solvent rather than to variable traces oi oxyEen, since the values observed were readily reproducible, and essentially the same results were obtained whether the "vacuum" was produced with an oil pump, with a water pump or whether the oxygen was displaced by prolonged bubbling of nitrogen through the solution. Further evidence for solvent effects per se was the observation that the position and character of the fluorescence bauds of benzpyrene vaned with the solvent both in air and in vacuo. Oxygen reduced the intensity of the bands without, however, altering either their position or discreteness (Fig. 2). Oxygen failed to alter either the intensity or the character of the absorption spectrum of benzpyrene. Other Inhibitors and the Fluorescence of Hydrocarbons.-Sulfur, like oxygen, inhibited the fluorescence of benzpyrene i n solution. Quenching, which amounted to 21.7'%, was observed with saturated solutions of rhombic sulfur in petroleum ether in vacuo; continued illumination resulted in the appearance of a yellow precipitate, which was also noted upon irradiation in the absence of benzpyrene. Nitrobenzene and tetranitromethane, previously reported as

Aug., 1943

THE EFFECTOF OXYGEN

ON THE

FLUORESCENCE OF CERTAIN HYDROCARBONS1543

fluorescence inhibitors6 in aerated solvents, also inhibited the fluorescence of benzpyrene in vacuo. Both in air and in vacuo the quenching produced by very small amounts of tetranitromethane was markedly and quickly enhanced by short exposures to ultraviolet light. Apparently the inhibition of fluorescence by these substances was a reversible process. Solutions exhibiting quenched fluorescence were passed through a column of aluminum oxide, and under ultraviolet light a fluorescent band of hydrocarbon appeared in each case when the column was washed with solvent. In previous experiments6 the addition of water to a solution of hydrocarbon in a miscible aerated solvent considerably enhanced the fluorescence of the solution. At least part of this effect was due to the lowered solubility of oxygen in the aqueous solvent. In vacuo, however, added water diminished the fluorescence somewhat; e. g., an inhibition of 1.6% was observed for benzpyrene in dioxane containing 10% water. In contrast to these various inhibitors, naphthacene consist ently failed to inhibit the fluorescence of these hydrocarbons in liquid solution whether in the presence of air7or in vacuo. Other Gases and the Fluorescence of Hydrocarbons.-Bowen and Norton3 first showed t h a t nitrogen had no effect on the fluorescence of certain hydrocarbons and that the quenching of fluorescence by oxygen could be easily and completely reversed by displacement of this gas from solution with a stream of pure nitrogen. In the Fresent experiments the quenching power of a series of gases was tested. Purified samples of nitrogen, hydrogen, carbon dioxide, carbon monoxide, hydrogen sulfide, sulfur dioxide, hydrogen chloride, ammonia, methylamine, and trimethylamine a t atmospheric pressure were passed through evacuated solutions of benzpyrene in petroleum ether or ethanol until saturation occurred and an oxygen-free atmosphere was assured. The maintenance of the fluorescence intensity observed in vacuo was taken as evidence of non-inhibition. Inhibition occurred only with hydrogen chloride, trimethylamine, and sulfur dioxide. The hydrogen chloride was freed of oxygen with chromous chloride and dried with coned. sulfuric acid. A consistent and reversible inhibition of fluorescence to the extent of 12% was observed a t 740 mm. in a solution of 0.02 p g . benzpyrene per cc. of petroleum

ether. Pure trimethylamine hydrochloride was preparedg and the free base liberated with excess sodium hydroxide. Saturation a t 740 mm. produced an inhibition of 96.4% in a solution of 0.1 pg. of benzpyrene per cc. of petroleum ether. Sulfur dioxide under the same conditions, completely and reversibly inhibited the fluorescence of benzpyrene, and fluorescence could only be observed a t relatively low values of Pso,. A t a partial pressure of 149 mm. sulfur dioxide was 25 times as potent as oxygen in quenching the fluorescence of benzpyrene ; a t higher pressures the ratio increased until a t 740 mm. no fluorescence was observed in a solution of 10.0 pg. of benzpyrene per cc. of petroleum ether. This concentration of hydrocarbon would yield a calculated 30,000 ammeter units of fluorescence in vacuo. Inhibitors and the Fluorescence of Tissue Extracts.-In a fluorometric determination of benzpyrene in the tissues of animals developing cancer, the effect of oxygen on the fluorescence of the natural substances from the tissues could become as important on the net result as the effect of oxygen on benzpyrene itself. Accordingly the non-saponifiable fraction from normal mouse tissue was prepared and its fluorescence in petroleum ether and in pyridine observed a t various partial pressures of oxygen. The fluorescent materials in the tissue extracts proved to be much less sensitive to dissolved oxygen than the hydrocarbons. Ratios of fluorescence intensities in air and in vacuo ranged from 1: 1.08-1.13 for pyridine, and 1: 1.40-1.69 for petroleum ether (Table 111). Corresponding ratios for the hydrocarbons averaged approximately 1:2 in pyridine and 1:4 in petroleum ether. It has previously been noted that the fluorescence of the tissue substances is relatively insensitive to changes in solvent. Theory of Fluorescence Quenching by Oxygen.-The partial pressure of oxygen necessary to give a molecular concentration of dissolved oxygen in petroleum ether equivalent to that of 0.1 pg. of benzpyrene per cc., as in our experiments, can be calculated to be only 3 x mm., a pressure well within our experimental error. It is doubtful whether this amount of oxygen affects fluorescence a t all. Conversely, it can be calculated that petroleum ether in equilibrium (9) Adams and Marvel, “Organic Syntheses,” John Wiley and Sons, Inc , New York. N Y , 1941. Coll Vol 1,p ,531

j.A.

15.14

THEEFFECT O F OXYGEN Tissue

Wet weight equivalents of extract aliquot,b g.

ON THE

MILLER AND

c. A.

VOl. 65

BAUMA“

TABLE111 FLUORESCENCE O F NON-SAPONIFIABLE MATTER‘ FROM MOUSETISSVS

-Petroleum etherAir Vacuum Ratio

Ammeter Units,c Wet weight equivalents

--PyridineAir

Vacuum

Ratio

Liver 2.4 25.1 35.5 1.41 1 . 2 g. 15.8 17.3 1.09 Organs i.8 20.1 34.0 1.69 0 . 9 g. 13.2 14.7 1.11 GI Tract 2.6 14.0 19.5 1.40 1 . 3 g. 11.8 12.7 1.08 Carcass 30.4 35.1 ,523 1.58 10.2 g. 28.9 32.8 1.13 The material used had been stored in petroleum ether in the dark for two weeks. Fresh solutions yielded somewhat higher ratios, 1.8-1.9 in petroleum ether and 1.2-1.3 in pyridine. Equivalent to petroleum ether extract of saponified Defined in Table 1. tissue from one 27 g. adult albino mouse.

with air a t 25’ and 740 rnm. dissolves approximately 10,000times as many molecules of oxygen as there are of benzpyrene a t this concentration. Under these conditions 15.5% of the original fluorescence still remains. Thus it follows that a t any instant only a small fraction of the total oxygen in the solvent can be actually reacting with the hydrocarbon. Complex formation, however transient, is often assumed t o be a basic step in the quenching of fluorescence, and the theory of its mechanism has been explored in some detail.1° Bowen and Williams* have concluded that for certain hydrocarbons fluorescence quenching by oxygen and photo-oxidation are similar but alternative processes, and may involve a dissociable collision complex of hydrocarbon and oxygen. WeilMalherbe and Weiss2 have suggested that the quenching of hydrocarbon fluorescence by oxygen may be represented by a chemical reaction in which an electron is transferred from the excited hydrocarbon to the oxygen molecule of the complex. However, whether the essential quenching reaction by oxygen involves a complex or a collision does not appear to have been determined. The hyperbolic relationship existing between the fluorescence intensity of benzpyrene and the partial pressure of oxygen (Fig. 1) may be explained in a t least two ways. On the basis of mass action i t may be assumed that in the quenching of fluorescence a non-fluorescent complex xHC-yO2 is formed. Since the observed quenching is easily and completely reversible, it would follow that this complex is easily dissociated as in xHC

+ yOa

xHC-~O.

The equilibrium constant of this reaction is expressed by K = pHCrYoz2 (HCP (Os)>

(1)

(IO) Bowen, Weiss and others A general discussion of luminerience held by the Faraday Society in Sept., 1938 Part I . Lumiiieqcence of Liquids and Vapours, T , a n ~Favadav Soc , Sb, 15 (1939)

If the concentration of hydrocarbon is n moles per liter and a = moles of &HC-yO*per liter

then n

- ax

= moles of free H C per liter

According to the assumption that the fluorescence of a solution is proportional to the concentration of free hydrocarbon, the fluorescence F = kl(HC) = kl(n

- ax)

By means of these values, and the Henry’s law equivalent, ( 0 2 ) = P o , , equation 1 can be reduced to .vK(kPO,)y1”

+

kl-.-l

F

- nklx

= 0

(2)

a general equation relating F and PO,in terms of the constants as defined. For the simplest complex, HC-02, x and y both equal 1, and equation 2 becomes (3)

This is similar in form both to the Stern-Volmer equationll for the quenching by foreign gases of the fluorescence of vapors at low pressures, and also to the equation used by Bowen and NortonY to express the dependence of the “fluorescence efficiency” (quanta emitted/quanta absorbed) of hydrocarbons upon the concentrations of such quenchers as dissolved oxygen or the hydrocarbon itself a t high concentrations. It is not necessary to know k and K separately for a given solvent, since

an expression in terms of observable and assumed quantities. Thus it is possible to calculate the F vs. Po, curve for any assumed complex of a given hydrocarbon and oxygen. All such calculated curves must pass through two experimentally determined points. One convenient point is ( F = kl, Po, = 0),characteristic of the (11 ) Hirschlaff, “Fluorescence and Phosphorescence.” Chemical Publishing C o , I n r . . New York. S. U , 1842, p. 18.

THEEFFECTOF OXYGEN ON

h g . , 1943

THEINTENSITY

OF

THE

FLUORESCENCE OF CERTAIN HYDROCARBONS1545

TABLE IV FLUORESCENCE OF BENZPYRENE I N VARIOUS SOLVENTS FOR DIFFERENTPARTIAL OXYGEN;OBSERVED AND CALCULATED'

PRESSURES O F

Intensity of fluorescence of 0.1 pg. BP/cc., ammeter units* mm.

Pol,

Cyclohexane Obs. Calcd.

0 21 42 63 116 155 310 466 742

304 226 190 162 120 95.5 57.8 39.4 21.0

(304)' 235 204 170 115 (95.5) 56.5 40.2 27.4

Obs.

Methyl cellosolve Calcd.

314

... ... ... 225 209 156 125 85.5

a All calculated for the complex BP.02. stants.

Tetrahydrofur-

Obs.

(314)

Pyridine Calcd.

301 281 ... 253 ... 238 228 200 (209) 175 130 157 125 101 92.4 72.0 Defined in Table I.

...

'

(301) 274 252 233 196 (175) 123 95.0 67.8

furyl alcoLui Obs. Calcd.

334

(334)

..

... ... ...

..

.,

..

.,,

286 250 220 178

(286) 247 219 178

Tricaproin

0 2 pg. BP/cc. Obs. Calcd.

PO2

0 79 157 373 447 747

36.3 32.7 30.9 24.5 23.8 18.0

(36.3) 33.3 (30.9) 25.6 24.2 19.8

Figures in parentheses were used to calculate con-

TABLE V OBSERVEDAND CALCULATED FLUORESCENCE INTENSITIES OF HYDROCARBONS FOR COMPLEX HC02 Fluorescence, Ammeter Units." Hydrocarbon, pg./cc. petroleum ether Poa,

mm.

M C , 0.1 pg. Calcd.

Obs.

DMBA, 0.1 pg. Calcd.

Obs.

DBA, 2.0 pg. Obs. Calcd.

BA, 1.0 pg. Calcd.

Obs.

A, 0.5 pg. Calcd.

Obs.

56.0 (56.0) 47.9 (47.9) 30.0 0 (30.0) 66.8 (66.8) 51.0 31.0 21.3 37.1 36.9 33.9 22.3 42.0 41.9 21 47.9 22.4 14.0 14.8 22.7 21.9 21.5 25.3 24.1 63 42.0 14.7 14.5 14.4 9.50 10.4 14.7 16.5 15.7 116 35.2 (11.6) 11.9 (11.9) 8.50 (8.50) 12.5 (12.5) 155 11.6 33.0 6.26 6.47 7.80 6.80 4.34 4.96 310' 6.64 6.89 23.8 4.59 4.40 4.48 4.74 3.04 3.49 4.55 466* 4.74 18.6 2.74 2.58 2.90 3.10 1.69 742b 2.29 3.06 2.71 11.6 a Defined in Table I. More concentrated solutions were used to obtain readings a t these pressures.

'

free hydrocarbon in the oxygen-free solvent. Another is observed fluorescence in air. These numerical values plus trial values of x and y serve for the determination of the KkY value, and the fluorescence a t any other partial pressure of oxygen can then be calculated. Figure 1 shows the calculated curves of various possible complexes of benzpyrene with oxygen in petroleum ether, as well as the observed fluorescence a t various pressures of oxygen. The curve obtained experimentally corresponds closely to the calculated curve for the complex BP-02; it does not correspond to any of the curves for the other possible simple complexes of B P and 0 2 . I n Table IV data are given for observed and calculated curves relating the partial pressure of oxygen to the fluorescence of benzpyrene in five other solvents including the triglyceride tricaproin. Experiments also have been performed with benzpyrene in purified cottonseed oil and in lard. The fluorescent materials in the crude fat were eliminated by treatment with tetranitromethane, after which colored materials were adsorbed on charcoal. I n Table V the results

(51.0) 47.5 41.7 36.3 (33.0) 24.4 19.3 14.1

of similar experiments are presented for each of the other hydrocarbons in petroleum ether. I n every case the agreement between the observed curves and the calculated curves for the complex HC-02 was sufficiently close to exclude other possible combinations. This agreement, however, does not necessarily prove that a combination between hydrocarbon and oxygen actually exists as such. It may merely express the probability. of bimolecular collisions, and an equation identical with equation 3 has been derived from a collision hypothesis.12 An attempt was made to distinguish between the mechanisms implicit in the two derivations by measuring the volume of air confined above pyridine before and after solid benzpyrene was dissolved in the liquid. The experiment was performed in a Barcroft respirometer, and the amounts of hydrocarbon dissolved were 10.16 and 0.68 mg. per 5 cc. Since 58% of the fluorescence of the hydrocarbon is quenched in pyridine a t the Po, of air, 526 and 35 ~ 1 of. oxygen, respectively, would have been absorbed if, in (12) M. J. Johnson,personal communication.

I540

J . A. MILLERAND C. A. BAUMANP;

quenching, the hydrocarbon had combined with the calculated amount of oxygen firmly enough to remove it efiectively from solution. However, no gaseous uptake was observed. Since no errors in combustion have been reported, it may be assumed that no oxygen was combined with the benzpyrene in the solid state. Hence within the limitations of the experiment, no evidence of firm combination was observed, and the collision hypothesis would appear to be favored if an essential difference actually exists between a collision and a short-lived complex. But, however interpreted, the data on the quenching of fluorescence by oxygen indicate that benzpyrene sotnehow reacts with dissolved oxygen, and that the reaction is not only regular and reproducible but even predictable from fundamental consider:itions. The results deny the popular irnprcssion of the capriciousness of fluorometric measurements. From a physiological point oi view it is of interest that the rccactiori between dissolved oxygen and the carcinogenic: hydrocx-bon proceeds in glycerides or in natural fats much as in any other solvent. U'hile the significance of the rcactioii is as yet obscure, it is possible that i t may be involved in the czrcinogenic process itself. Another possibility is that the reaction is the first step in a sequence by which the hydrocarbons are rendered biologically inactive.

Vol. 65

In air the intensities usually ranged from onehalf to one-sixth of those observed in vacuo (absence of all gas but solvent vapor). Most of the difference was due to dissolved oxygen, although in some cases the degree of quenching did not parallel oxygen solubility. In the presence of oxygen the fluorescence bands of benzpyrene were diminished in intensity, but both the fluorescence and abswption spectra of benzpyrene were qualitatively the same in air and in taczio. 2. Sulfur dioxide was many times as efiective an inhibitor oi fluorescence as oxygen ; hydrogen cliloridc and trimethylamine were less effective. The inhibition was reversible by changing the partial pressure of the gas. Sulfur, nitrob-nzene, and tetranitromethane inhibited the fluorescence of benzpyrene solutions both in air and in vacuo 'This inhibition could be reversed by passing the quenched solution through a column of aluminum oxide. Sitrogen, hydrogen, carbon dioxide, hydrogen sulfide, methylamine, and ammonia failed to afiect the intensity of fluorescence of benzpyrene. :{. The fluorescence intensities of the hydrocarbons varied hyperbolically with the partial pressure of oxygen. Identical equations relating these two quantities were derived on the alternate assumptions that a non-fluorescent dissociable complex formed between oxygen and the hydrocarbon or that the quenching of fluorescence by oxygen was a collision phenomenon. ObSummary qerved values agreed well with calculated values I . The intensity of fluorescence was measured from the derived equations. for S,-l-benzpyrene, 20-methylcholanthrene, 9,lO4. 'The fluorescence of the non-saponifiable dimethyl-l,2-benzanthracene, 1,2,3,B-dibenzan- matter from mouse tissue did not vary greatly thracene, lj2-benzanthracene and anthracene, with Po, either in petroleum ether or in pyridine. h in se\-eral solvents in air and in m m o . I\.f..4DISOX, WISCONSIN RECEIVED MARCH11, 1943